remove duplicates based on a filter

Hi IF ,

 

How about this:

 

Before:

 

After:

 

 

Here the code in Power Query M that you can paste into the advanced editor (if you do not know, how to exactly do this, please check out this quick walkthrough)

 

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("nZExDsIwDEWvgjJXlr+dUBiBuVL3qve/BnFoTEjp0sGO7OTJ/zvLEqJG5nQLQ5iRE0BIJCySi8mSqLVzvOwU4vFzvQ4VlhaODpeOtDBI7j2M/3DlLB45Lg3CjjBBHYk56QGjxaDsDBZv6DReO43lqcNfjaNt55fFjoWz3PuLaYOfnVhu5x0YrL+BMwPPbHQTub4B", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [ID = _t, Name = _t, DateStart = _t, Type = _t, Days = _t, Qty = _t, Classification = _t, DateClose = _t]),
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"ID", Int64.Type}, {"Name", type text}, {"DateStart", type text}, {"Type", type text}, {"Days", Int64.Type}, {"Qty", Int64.Type}, {"Classification", type text}, {"DateClose", type text}}),
    #"Grouped Rows" = Table.Group(#"Changed Type", {"ID"}, {{"AllRows", each _, type table [ID=nullable number, Name=nullable text, DateStart=nullable text, Type=nullable text, Days=nullable number, Qty=nullable number, Classification=nullable text, DateClose=nullable text]}, {"CountRows", each Table.RowCount(_), Int64.Type}}),
    #"Filtered Rows 1" = Table.SelectRows(#"Grouped Rows", each ([CountRows] > 1)),
    #"Expanded AllRows 1" = Table.ExpandTableColumn(#"Filtered Rows 1", "AllRows", {"Name", "DateStart", "Type", "Days", "Qty", "Classification", "DateClose"}, {"AllRows.Name", "AllRows.DateStart", "AllRows.Type", "AllRows.Days", "AllRows.Qty", "AllRows.Classification", "AllRows.DateClose"}),
    #"Filtered Rows 2" = Table.SelectRows(#"Expanded AllRows 1", each ([AllRows.Classification] = "B")),
    #"Filtered Rows 3" = Table.SelectRows(#"Grouped Rows", each ([CountRows] = 1)),
    #"Expanded AllRows 2" = Table.ExpandTableColumn(#"Filtered Rows 3", "AllRows", {"Name", "DateStart", "Type", "Days", "Qty", "Classification", "DateClose"}, {"AllRows.Name", "AllRows.DateStart", "AllRows.Type", "AllRows.Days", "AllRows.Qty", "AllRows.Classification", "AllRows.DateClose"}),
    #"Appended Query" = Table.Combine({#"Expanded AllRows 2", #"Filtered Rows 2"}),
    #"Renamed Columns" = Table.RenameColumns(#"Appended Query",{{"AllRows.Name", "Name"}, {"AllRows.DateStart", "DateStart"}, {"AllRows.Type", "Type"}, {"AllRows.Days", "Days"}, {"AllRows.Qty", "Qty"}, {"AllRows.Classification", "Classification"}, {"AllRows.DateClose", "DateClose"}}),
    #"Removed Columns" = Table.RemoveColumns(#"Renamed Columns",{"CountRows"})
in
    #"Removed Columns"

 

The idea is to do a group by and filter for the IDs that have more than one row. For these rows, we filter for classification = B. Then we append that resultset with all rows that just have one row in the group by. In the end it's a bit of cleaning like renaming columns and removing the count (support) column.

 

I'd suggest to paste in the M code in a new query and check out the steps on the right hand side.

 

Let me know if you have questions 🙂

 

/Tom
https://www.tackytech.blog/
https://www.instagram.com/tackytechtom/